PERFORMANCE-LEVEL SEISMIC MOTION HAZARD ANALYSIS METHOD BASED ON THREE-LAYER DATASET NEURAL NETWORK

摘要

The invention relates to an anti-seismic technique analysis method, in particular to a performance-level seismic motion hazard analysis method based on a three-layer dataset neural network. The method comprises the following steps: (S1) extracting seismic motion data and denoising the data; (S2) extracting feature data from the data, and carrying out initialization; (S3) generating a training set, an interval set and a test set; (S4) training a multi-layer neural network based on the training set; (S5) training output values of the neural network based on the interval set, and calculating a mean and a standard deviation of relative errors of the output values; (S6) training the neural network based on the test set to determine output values, and calculating a magnitude interval based on an interval confidence; (S7) carrying out probability probabilistic seismic hazard analysis to determine an annual exceeding probability and a return period of a performance seismic motion; and (S8) determining a magnitude and an epicentral distance that reach the performance-level seismic motion based on the performance seismic motion and consistent probability. A novel neural network training method is used to predict the seismic motion attenuation relation, thus improving the universality and flexibility of the attenuation relation.